This program enables you to simulate correlated multivariate binary data according to the algorithm of Emrich and Piedmonte (1991). The program is from Wicklin (2013). The attached file is RandMVBinary.zip, which contains two files:
- RandMVBinary.sas defines the RANDMVBINARY function, along with some helper functions. Save this file so that it can included into a SAS session. For example, save it to C:\Users\<userid>\Documents\My SAS Files\RandMVBinary.sas
- SimMVBinary.sas is a "driver" program that contains three examples of generating binary data. It shows how to load and call the RANDMVBINARY function.
Note that you should %INCLUDE the RandMVBinary.sas file before calling the IML procedure. Within the procedure, use the LOAD statement to load the RANDMVBINARY function before you call it.
The RANDMVBINARY function uses features of SAS/IML 12.1, which was released as part of SAS 9.3m2. If you are running an earlier version of SAS, call the function RANDMVBINARY93, which has the same syntax but is slightly less efficient.
References:
Emrich, L.J. and Piedmonte, M. R., 1991, "A Method for Generating High-Dimensional Multivariate Binary Variables," The American Statistician, 45, p. 302--304
Wicklin, R., 2013, Simulating Data with SAS, SAS Institute Inc., Cary NC, pp. 154--157.
